• 1  Decourt – Tziaphalias (2015). For the date, see CGRN 225. Abbreviations for Greek inscriptions fol (...)
• 2  There have been three subsequent editions: Bouchon – Decourt (2017); SEG 65.376; CGRN 225. For stu (...)

1In 2015, the inscription from Marmarini, an extraordinary Greek inscription with cult regulations for an unknown sanctuary in Thessaly (ca. 225–150 BCE), was first edited.1 Ten years on, this unique document has lost none of its fascination; its distinctive mixture of Greek and Near Eastern religious features is still a mystery and remains difficult to explain.2

• 3  For brief comments, see Bouchon – Decourt (2017), p. 178–179; CGRN 225, commentary § 6 (lines 48–5 (...)
• 4  See Decourt – Tziaphalias (2015).
• 5  See Bouchon – Decourt (2017).
• 6  See CGRN 225, with commentary on § 6 (lines 48–54).
• 7  CGRN 225, commentary on § 6 (lines 48–54).

2Among its various cult regulations, the Marmarini inscription also contains a rubric on oath-taking. This rubric, which covers lines 48–55 on face A, has not yet received much attention and awaits a detailed discussion.3 It is one of the “youngest” parts of the text, so to speak, whose content has been deciphered only gradually. In the editio princeps (2015), J.-C. Decourt and A. Tziaphalias were able to identify a few clusters of words in this section, but the readings were yet too raw and fragmentary to even be certain about the general topic.4 The improved text of the follow-up edition by R. Bouchon and J.-C. Decourt (2017) clarified that the section was indeed about oaths and also revealed several details, but the readings were still very fragmentary and uncertain.5 The edition by J.-M. Carbon and C.V. Crowther, which was published as CGRN 225 (online 2020), offered yet again several improved and new readings; a clearer picture emerged about the content of the section.6 Even though the text of lines A 48–55 is still fragmentary and has been described as “considerably obscure, even where legible”,7 the current text provides an excellent starting point for further improvements and a systematic analysis.

• 8  On oaths in ancient Greece, see Latte (1920); Plescia (1970); Graf (2005), p. 237–246; Sommerstein(...)
• 9  For “full” descriptions of oaths, see e.g. SEG 57.576 (Dikaia, 365–359 BCE, political); I.Iasos 2. (...)
• 10  See e.g. IG XII 4, 132.126–128 (ca. 300 BCE); 152.9 and 32 (ca. 208 or shortly after); Plescia (19 (...)

3As is hinted at in CGRN 225, the obscurity of this section is due both to the fragmentary state of the text and its content. The oath regulations differ in almost all respects from the way in which oaths are described in Greek sources.8 Normally, Greek inscriptions and literary texts put much emphasis on the oath formula and the invocation of the gods; they make it clear who the oath-takers are, in which situation an oath is sworn and for which reason; and they include a conditional curse.9 When an oath ritual is mentioned, it is normally a gesture that revolves around animal sacrifice and libation, the two most common rituals in Greek religion.10 The Marmarini inscription, by contrast, refers to five different oaths in as many as six and a half lines, but does not mention the wording of any of these oaths. What is more, it prescribes ritual actions such as the grasping of a “weapon of the goddess” (A 50) and the act of drinking “from the altar” and “from the fish” (A 51–52). None of these actions is attested as an oath gesture in extant Greek texts. Most strikingly, the text defines different monetary values as reasons for taking an oath and stipulates payments for the oath-takers. The section on oath-taking thus raises both specific questions about the individual rules and more general questions about the meaning and function of the oaths and their cultural background.

• 11  In the Athenian legal practice of the classical period, the role of evidentiary oaths (both oppone (...)

5In addition to several smaller observations, the article makes three major arguments. The first argument concerns the structure of the section on oaths. It is argued that the individual rules are interconnected and form a coherent whole that is meant to be read accordingly. The second and central argument concerns the meaning and purpose of the regulations. Distinguishing two kinds of oaths in lines A 48–55, the article develops the hypothesis that the primary function of the rubric is to regulate the use of oaths in a judiciary context. In particular, it is argued that the sanctuary exercised a temple jurisdiction, offering to its visitors the circumscribed “legal service” of a small claims court where personal disputes were decided on the basis of evidentiary oaths. Depending on certain conditions, the oaths took different ritual forms. The third major argument concerns the socio-cultural background of this judiciary oath practice in the sanctuary. The distinctive use of evidentiary oaths in the temple strongly points to legal practices in the ancient Near East as a cultural sphere of influence.11 This includes the mysterious actions of grasping a “weapon” of the goddess and drinking “from the altar” and “from the fish”. Significantly, these actions accord with gestures known from Near Eastern legal practice and oath-taking in sanctuaries.

• 12  At the end of each line, I have included, in round brackets, the total number of letters that have (...)

6I start by giving the text and translation of lines A 48–55 as presented in CGRN 225.12

• 13  This calculation also accords with the average number of letters in lines A 20–38: the average is (...)

8The rubric starts with the first conditional clause in line A 48. Though it is not obvious in the text presented in CGRN 225, line 48 is the most fragmentary line of the entire section. In CGRN 225, a lacuna is indicated between περ̣ί̣ and ὀβ̣[ο]λόν, with a lone majuscule alpha dividing the lacuna into two unequal parts: in addition to a smaller space for just one letter between περ̣ί and the lone alpha, there is a larger gap between this alpha and ὀβ̣[ο]λόν. The editors do not offer an estimate of the size of this gap, but it is clear that it is substantial. The text in line A 48 fills the entire width of the stele, starting on the left margin with ὅρκον and ending on the right edge with τιθ̣έ̣τ̣ω̣. According to my own estimate, the larger gap offers space for approximately 20–22 letters. Given that 35 letters have been read or identified in this line, this estimate fits in well with the expected total number of letters of a full line in this part of the stone, which is between 55–61.13

• 14  The use of εἰ also seems unlikely in view of the other conditional clauses in the inscription. In (...)
• 15  See Kühner – Gerth (1898), p. 41–42, § 354, Anmerk. 2c. It is curious that the expected ἦι would f (...)
• 16  There are several indefinite clauses construed with ὅ τι ἄν (or the like) in the inscription. An e (...)
• 17  Bouchon – Decourt (2017) read θ̣υσάτω̣, which already correctly identifies the grammatical form bu (...)

12Before discussing how the conditional clause in line A 48 may be connected with the participle in line A 49 (π̣[ε]ριλ̣αβὼν ὅπλ̣ο̣ν τῆς θεοῦ ὅ τι ἂμ̣ βούληται), it is necessary to address the lacuna in line A 49 (ὅ τι ἂμ̣ βούληται [....^{c. 8}....· εἰ?] δὲ μνᾶς ἄξ̣ιον,^v). This lacuna roughly occupies the middle part of the line where the inscription has worn away most. Carbon and Crowther estimate that the empty space is large enough for circa ten letters, including the two letters of the conjunction εἰ, which they tentatively suggest. The grammatical markers after the lacuna, namely the particle δέ before μνᾶς ἄξ̣ιον (line 49) and the subsequent double stipulation that the oath-taker should pay three obols and take or grasp something (line 50), make it indeed certain that a new conditional clause begins in line A 49 and that a conjunction is missing. But it appears that instead of εἰ, the conjunction is ἐάν.14 In the space at the end of the lacuna, the distinctive shapes of an alpha followed by nu can be recognized (with the nu a little less clear than the alpha) and, before these letters, traces that look more like epsilon than another letter (see Fig. 2). The ellipsis of εἶναι in the subjunctive may seem surprising, but there is one more case of such an ellipsis in an ἐάν clause in the inscription (ἂν δὲ καταβε̣|βρωμένος, θερα̣π̣εύειν, A 28–29), and it is also attested elsewhere.15 While it is thus clear that a new conditional clause begins with ἐὰν δέ in line A 49, the question remains what was written in the space before the conjunction. After the participle π[ε] ριλ̣αβὼν ὅπλ̣ο̣ν τῆς θεοῦ ὅ τι ἂμ̣ βούληται (A 49), which specifies that the oath-taker is allowed to grasp “any weapon of the goddess he wishes”, one could expect an “except” phrase (πλήν τινος) that identifies an implement of the goddess that is off limits for the oath-taker.16 However, the images again suggest something else. Right after the last letter of βούληται in A 49, the characteristic circle of an omicron is visible. This omicron is followed by the distinct traces of mu and, in this sequence, by shapes of the following letters: the left vertical and parts of the diagonal of nu; the traces of the two arms of upsilon (the arms seem relatively narrow); a vertical with parts of three horizontals, which suggests epsilon; and the traces of the cross of tau (see Fig. 3). I suggest, therefore, to read ὀμνυέ̣τ̣[ω].17 We receive the following text in line A 49:

• 18  The same syntactical construction of a participle in the aorist, followed by the imperative of the (...)

15As regards the actual oath, the first conditional (line 48–49) offers the following details: the protasis introduces the rubric, referring to the situation of taking an oath in the sanctuary in general (a). The apodosis provides the information that in this case, two things are required: a ritual action (b) and a payment (c). Then, the sentence specifies the object used in the ritual action (d) and explains the sequence of actions (e).18 The additional detail offered by (d) is important, because the information given by (b) focuses on the need to perform a particular ritual action, leaving it open which “weapon” of the goddess may be touched.

• 19  This may also explain the use of the simplex λαμβάνω instead of the compound περιλαμβάνω, which ha (...)

19Importantly, these instructions involve the same ritual action as the “general oath” in lines A 48–49: the oath-taker has to touch an object described as a ὅπλον̣ or “weapon” belonging to the goddess. In this case, however, the definite article τό is used, because it has already been explained that the oath-taker may grasp “any” weapon of the goddess.19 What is different in this oath is the amount of the required payment. Instead of one obol, the oath-taker has to pay three obols.

• 20  The construction ἐὰν δέ τις βούληται also appears in lines B 7–8, 23, 34, 44–45 48–49, 57, 60 (var (...)
• 21  The form ἁγνεύσηι is surprising, not only because one expects another imperative of the third pers (...)
• 22  Fines for not taking an oath are usually high. In the sacred law from Andania (IG V 1, 1390.6–7), (...)
• 23  For these reasons, other aspects of oath-taking, such as the task to administer an oath (ὁρκίζειν, (...)
• 24  See also CGRN 225, commentary § 6 (lines 48–54) where the connection between special purity regula (...)

22The fourth conditional clause, which starts in line A 51 and extends almost to the end of line 52 (ἐὰν δέ τι̣[ς] Ν[.]Ο̣Ρ̣Μ̣[..]Ι̣Ν̣ β̣ο̣[ύ]λ̣η̣τ̣αι, ἀ̣[π]ὸ τοῦ ἰ̣χθ̣ύος π̣ι|ν̣έτω καὶ τιθ̣έτω δραχμὰν αἰγιναίαν κα̣ὶ̣ [ἁ]γν̣εύσηι ἕως ἂν ζώη̣ι̣), would be complete, if it were not for a fragmentary series of six disjointed letters, all of which are marked as uncertain readings in CGRN 225 except for the initial nu. The basic syntactical structure of the protasis (ἐὰν δέ τι̣[ς]… β̣ο̣[ύ]λ̣η̣τ̣αι) makes it practically certain that we are looking for an infinitive. The expected sense would be that someone wishes “to do or be something”.20 The crucial question is what exactly this action might be. We may start by considering the requirements in the apodosis. Two things are striking: the unprecedented ritual instruction that the oath-taker “should drink from the fish” and the additional statement that he “will be/keep pure as long as he lives” ([ἁ]γν̣εύσηι ἕως ἂν ζώη̣ι̣).21 This obligation is an extraordinarily strict purity requirement, and it must have everything to do with the missing word in the protasis. Given the oath context, one may expect that the condition concerns the refusal to swear an oath. Oath regulations often include such a scenario. However, if this were the case, it would be difficult to explain why lifelong purity is mentioned as a requirement in the apodosis. What is usually stipulated in these cases is a large fine, sometimes 1000 drachmae.22 In any case, a reading like μὴ ὀμνύειν, seems too long for the space of the lacuna and may also be excluded because of the initial nu, which seems certain.23 Instead, the strict purity requirement may be indicative of a cultic function.24

• 26  On the role of the neokoros in Greek religion, see Georgoudi (2005), p. 57–58; see also Ricl (2003 (...)
• 27  In these passages, the noun νεωκόρος is used. The verb νεωκορεῖν is attested a few times in Greek (...)
• 28  There are textual and syntactical issues in these lines, but the group of sacred personnel is cert (...)
• 29  For the entry regulations in the Marmarini inscription, see CGRN 225, commentary, § 1 (lines B1–6) (...)
• 30  See CGRN 225, commentary § 2 (lines B7–12); Zerhoch (2025), p. 164.

25We know that the sanctuary had a νεωκόρος or “temple warden”.26 A temple warden is mentioned in lines A 7, B 4 and B 33 together with other sacred personnel.27 In lines A 4–8, it is stated that together with the priestess (μετὰ τῆς ἱερείας), the women called phoibatriai, the temple warden, and “anybody else of the initiated who volunteers” (τὰς φοιβατρίας καὶ τὸν νεωκόρον καὶ ἄν τις ἄλλος βούληται τῶν τετελεσμένων) are in charge of cleaning the sanctuary and of performing a sacrifice and a purification ritual before the major festival of the Eloulaia.28 In B 4 and B 33, the temple warden is mentioned as one of the officials in charge of purification rituals that need to be performed when a visitor does not fulfil certain entry restrictions:29 in B 4–5, it is stipulated that “the priestess or the neokoros or one of the women lifting the sacred objects should perform the purification” (καθαίρειν τὴν ἱέρειαν ἢ τὸν νεωκόρον̣ | ἢ τῶν τὰ ἱερὰ αἰρουσῶν τινα), if someone enters the temple (ναός) of the goddess without being initiated. In B 33–34, it is “the neokoros and whoever of the phoibatriai is present” who are charged with purifications (καθαίρειν δὲ τὸν νεωκόρον ταῦτα καὶ τῶν φοι|βατριῶν ἥτις ἄμ παρῆι). This case concerns the situation that a person has entered a specific area of the sanctuary—most likely the courtyard peristyle30—without fulfilling the appropriate hagneia rules (ἐὰν δέ τις εἰσέλθηι μὴ ἁγνεύσας τῶν προγεγραμμένων, … Β 28–29).

• 31  See Georgoudi (2005), p. 57–58.
• 32  For oaths taken by sacred personnel, see e.g. IGV 1, 1390.1–11 (Andania); SEG 47.1628.16– 24 (Eryt (...)
• 33  As has been noted in CGRN 225, commentary § 6, lines A 48–54, life-long purity of priests is not n (...)
• 34  For the dramatic function of purity in Euripides’ Ion, see Gibert (2019), p. 46–49. Interestingly, (...)
• 35  See SGDI II 1884 (170–157 BCE); 1912 and 1913 (both 156–151 BCE); 1701 and 1957 (both 150–140 BCE)
• 36  See e.g. IG XII 9, 906.2 (Chalkis, after 212 CE, ὁ διὰ βίου νεωκόρος); IGLS 3, 2, 1185.2–3 (Seleuc (...)
• 37  Though the text in lines A 7, B 4 and B 33 strongly suggests only one neokoros in the sanctuary, t (...)

26It seems that the neokoros took care of the daily organization of the cult, making sure, either together with or in place of the priestess and other sacred personnel, that the visitors would follow the required purity rules. In view of these tasks, the fourth conditional starts to reveal its meaning. The neokoros is usually regarded as a subordinate helper in the organization of a sanctuary. But it has also been argued that this view does not do full justice to the status and responsibilities attached to this office.31 The rule in the Marmarini inscription reflects both sides. On the one hand, it suggests that this office was principally open to anyone who wished to hold it. But on the other hand, the rule makes it clear that νεωκορεῖν involved cultic tasks that were of great significance for the sanctuary, such as purification rituals, which also included sacrificial actions. The fact that the person who wants to be a neokoros has to swear an oath (implied) and fulfil the hagneia requirements for the rest of his life highlights this significance. The rule informs the candidate not only about the oath-taking itself, namely the payment of a drachma and the performance of a special ritual gesture (drinking “from the fish”), but also about the obligation of a life constantly bound by purity regulations.32 Since Greek inscriptions usually speak about the conditions of priesthoods and do not specify those of the office of a temple warden, it is unknown whether such a strict obligation was common practice. But the emphasis on this aspect in the Marmarini inscription makes particular sense if the neokoros served for life and had to take care of the sanctuary and its visitors on a daily basis. In order to be able to perform priest-like functions and purifications under these conditions, the temple-warden had to be in a state of ritual purity himself.33 Interestingly, in Euripides’Ion, Ion, who is introduced as a boy living in the sanctuary of Apollo at Delphi and performing the tasks of a temple warden, deeply cares for the purity of the sanctuary, and in his monody, he expresses the wish to serve the god his entire life (εἴθ᾽ οὕτως αἰεὶ Φοίβωι | λατρεύων μὴ παυσαίμαν, | ἢ παυσαίμαν ἀγαθᾶι μοίραι, 151–153).34 Manumission documents from Delphi regularly mention the priests and the neokoros, by name, as witnesses of the transactions taking place in the sanctuary; a certain Menes, who appears in several of these inscriptions, may have been neokoros in Delphi for as long as thirty years.35 There is also evidence from honorary inscriptions of the Roman period that neokoroi did serve for life.36 The special oath instructions in the Marmarini inscription for a person who wishes to become a neokoros are thus indicative of the importance attached to the status of this office in the sanctuary.37 In addition to mentioning the purity requirement, the rule for the oath of the neokoros asks for a higher payment than in the other cases and for a particular oath gesture, namely to drink “from the fish”, which, as we will see, most likely refers to an action that is meant to bring the person in close contact with the anonymous goddess.

• 38  For purifications in the form of washing κατὰ/ἐκ κεφαλῆς, see Ehrenheim (2022), p. 41–48, who expl (...)
• 39  See CGRN 225, commentary § 7, lines A 54–55.
• 40  It is noteworthy that the sacrificial regulations on face B (34–73) include offerings in the form (...)
• 41  For the mention of sacrificial ἱερά in the context of oath-taking, see e.g. IG XII 4, 132.127– 128 (...)

27The fifth and final conditional clause in lines A 52–54 (ἐὰν δέ τι̣ς | περὶ μικροῦ ὀμνύηι, λουσάμενος κατὰ κεφα̣[λ]ῆς, ὀμνυέτω ἐν τῶ̣ι γνύθ̣ω̣ι | καὶ τιθέτω ἡμιωβέλιον) raises no textual questions. However, it offers again new details that are important for the understanding of the entire rubric. The rule states that if someone swears an oath “about something small” (περὶ μικροῦ), the person should, after performing a ritual washing “from head down” (λουσάμενος κατὰ κεφα̣[λ]ῆς), swear in a place called the “pit or hollow” (ἐν τῶι γνύθωι, 53) and pay half an obol.38 As in the case of the second conditional in lines A 49–50, the form of the oath and the amount of money to be paid are determined by the value of the cause for the oath. Though this rule does not specify a ritual oath gesture, the facts that it refers to a designated place where the oath is sworn and includes a purification implies a ritual. As the context makes clear, this ritual is different from the ritual gestures mentioned in the first four conditionals, namely the touching of a weapon of the goddess and the drinking from “the altar” or “the fish”. The inscription itself offers a clue as to its possible form. After the last conditional, there is one more sentence on face A. This sentence, which is separated from the fifth conditional clause by a punctuation mark in the form of a colon (: ), simply states: “The half obols should be paid to the man who did the slaughtering (of a sacrificial animal) and took out the sacred parts” (τελεῖν̣ {Ι̣ΩΙ} τῶι σφάξαντι κ̣α̣ι̣ τὰ ἱερὰ ἐξελόντι τὰ̣ ἡ̣μ̣[ι]|ωβέλια, A 54–55). In the commentary in CGRN 225, it is suggested that this rule may be independent of the oath regulations and refer to payments made to a μάγειρος or butcher for his work concerning the animal sacrifices performed in the sanctuary.39 But a stronger case may be made that it applies first of all to the last oath-regulation and, perhaps, by extension to animal sacrifices in general.40 Since the prescription for the payment of the half-obols in lines A 54–55 is made directly after a half obol is mentioned as an oath fee (A 54), it would be hard not to read it in the sense that the slaughterer receives it for his work in the context of the oath ritual. This is also supported by the grammar: the definite article τὰ̣ ἡ̣μ̣[ι]|ωβέλια (54–55) is used right after an unspecified ἡμιωβέλιον is mentioned as payment in the singular (54). If this consideration is correct, the fifth conditional clause implies that the oath-taking in the “pit or hollow” includes a sacrificial oath gesture. The reason for this specification could be that the sacrificial ἱερά played a significant part in the oath sworn in this location. When Greek inscriptions mention a sacrificial oath ritual, they often emphasize the role of ἱερά, specifying, for instance, that an oath is taken “over freshly burnt ἱερά”, or “when the ἱερά are burning”, that is, when the sacred portion is burned on the altar.41 The ritual in the place defined as a “pit or hollow” may involve this or another kind of ritual manipulation of the ἱερά (it does not need to involve a burnt offering); in any case, it would be a sacrificial action. Two things follow: first, the last sentence on face A forms an appendix to the oath-regulations; and second, the rubric on oath-taking is the last set of rules on this side of the stone.

• 42  For the interpretation of the spelling ἁγνεύσηι as future indicative, see above n. 21.

30As can be seen in Table 1, the rubric for oath-taking is organized around a sequence of five conditional clauses (protasis and apodosis). The protasis always consists of a construction of ἐάν + subjunctive, while the instructions in the apodosis are given in the imperative of the third person or, in a single case, in the future tense (A 52).42 Each conditional refers to a particular form of oath. For analytical purposes, I have designated the five oaths in Table 1 with the letters from A to E (col. 1). Table 1 shows not only the different conditions and requirements that are mentioned in the text, but also gives information that is implied; it brings out the correlations between the different kinds and levels of conditions, on the one hand, and the requirements, on the other. Significantly, there are two classes of conditions (cols. 3 and 4), and normally just one condition is given, either one of the first or one of the second class. Table 1 indicates, furthermore, that an oath is always defined by two requirements (cols. 5 and 6).

• 43  For the meaning of the phrase πίνω ἀπό τινος, see below n. 123.
• 44  For the question of the meaning of ὅπλον, see the discussion in section 3.
• 45  For this question and the implications, see the discussion in section 3.

34It is conspicuous that the text does not mention any oath formulas. Instead, the inscription shows a clear and distinct interest in another aspect of oaths: it emphasizes the role of ritual (col. 6). Table 1 shows that each conditional clause refers to a ritual gesture as a requirement for an oath. This emphasis on ritual suggests that the sign function of the oath gesture is considered crucial as a means to make the oaths effective. In particular, we can identify two significant aspects of the rituals: the action itself and the contact of the oath-taker with the goddess as is implied in the action. The fact that the different forms of ritual correspond with the different conditions and with the scaled fees of ½ obol, 1 obol, 3 obols and 1 drachma, also shows that the oaths were organized around a religious principle, namely how closely the oath-taker gets in contact with the main goddess of the sanctuary. Considering all factors and correlations, we may establish the following levels of proximity to the goddess: (1) the oath-taker seems furthest removed from the goddess in oath E, which involves a sacrificial action performed in the “pit or hollow” (ἐν τῶι γνύθ̣ω̣ι, A 53). (2) In the case of oath C (= A′/B′), which concerns a person who is not initiated, the ritual brings the oath-taker physically closer to the goddess, but the contact is still indirect: the oath-taker has to “drink from the altar”. The phrase “to drink from something” (πίνω ἀπό τινος) suggests that the person would go to the altar (of the goddess), take a drinking vessel that is placed there and drink from this vessel when swearing the oath.43 In this case, the contact with the goddess is mediated by the place (the altar) and its religious significance. (3) The ritual required for oaths A and B, which concern persons who are initiated, seems to bring the oath-taker in much more direct contact with the goddess. In this case, the oath-taker grasps a “weapon of the goddess”. Two scenarios seem likely: the ὅπλον may be either an actual weapon or an implement of the goddess considered a divine weapon, or it may be a weapon that was dedicated to the temple and became the possession of the goddess.44 (4) The ritual that provides the closest contact with the goddess is the mysterious drinking “from the fish” in oath D. Crucially, this object only makes sense in an oath gesture, if it has a religious significance and represents the goddess and her sphere of influence.45

• 46  Staatsverträge III 558 (F.Delphes III 1, 486). For discussions, see Haussoullier (1917); Cataldi ( (...)
• 47  Staatsverträge III 558, I, A.1.3–4, with commentary p. 343.
• 48  Staatsverträge III 558, I, A.1.7–10.
• 49  For the agronomoi, see Morrow (1960), p. 186–190; Schöpsdau (2003), p. 404–415.

37There is evidence that the value of a claim was a principle in ancient Greek legal practice. A fragmentary contract for legal help (Rechtshilfevertrag) between Delphi and Pellana dating to the first half of the third century BCE formulates rules for equal legal procedures in both cities concerning certain offences committed by citizens from the respective other city.46 The contract explicitly refers to the value of the claim: depending on whether the matter in question is worth less or more than five minas (500 drachmas), the case is tried before a smaller or larger group of judges.47 In addition, the document stipulates that judges and opponents take oaths while pouring libations over burnt offerings.48 A passage in Plato’s Laws is even more telling. In book 6 of the Laws, the discussion turns to the political organization of the rural parts of Magnesia. Plato introduces a committee of five ἀγρονόμοι or land-stewards.49 In addition to administrative functions, these magistrates have judicial authority (761d–e):

• 51  For the distinction between “the five” and “the twelve”, see Schöpsdau (2003), p. 404–407.
• 52  Interestingly, Plato states that the villagers have jurisdiction over the land-stewards in cases o (...)
• 53  For discussions, see Morrow (1960), p. 241–296; Schöpsdau (2003), p. 424–437. Significantly, the v (...)

38The law in this passage determines that the five land-stewards have the right to decide cases involving minor damages (τὰ μὲν σμικρά), while larger cases (τὰ δὲ μείζονα) involving values of up to three minas should be decided by a committee formed by the five and their twelve associates.51 Several aspects are relevant for our discussion of the oath section in the Marmarini inscription. First, both the Rechtshilfevertrag between Delphi and Pellana and the passage in Plato’s Laws suggest that the value of a claim was a common principle in legal procedures in ancient Greece. Second, the distinction between the two values which Plato makes for the personal disputes in the rural areas of Magnesia matches almost exactly the distinction the Marmarini inscription makes for the oaths. Both Plato and the Marmarini inscription define the lower value of the cause as something “small” and use as higher value the unit of a mina. The only difference is that the inscription sets the value at one mina, while Plato sets it somewhat higher at three minas.52 Third, the legislation of the ἀγρονόμοι represents a level of legal arbitration that is situated outside the “high” courts of the polis.53 However, there is also a major difference. The cases referred to in the inscription from Delphi and in Plato are not tried in a sanctuary. They are communal affairs.

• 54  See Lipsius (1915), p. 167–175, who points out that most of the courts in Athens must have been lo (...)
• 55  See e.g. IG XII 1, 152.2 (Cos, 208 BCE or shortly after); Aristot. Ath. Pol. 53, with Plescia (197 (...)
• 56  For sworn treaties/decrees involving oaths put up in a sanctuary, see e.g. IG II³ 1, 912.95–97 (26 (...)
• 57  IG XII 4, 356.4 (LSCG 175).
• 58  IG VI 1, 1390.1–11. For the dating, see Gawlinski (2012), p. 3–11; CGRN 222. Gawlinski (2012), p.  (...)
• 59  SEG 47.1628.16–24 (CGRN 98 fr. B); see also Dignas (2002), p. 29–40; Herrmann (2002), p. 157–172.
• 60  Though the text in lines 24–26 is very fragmentary, it seems likely that it stipulates a fine for (...)

39Ancient Greek law courts were not normally situated in sanctuaries or presided over by sacred personnel.54 Judicial oaths were usually sworn at a sacred stone or altar in a central public space like the agora.55 What was of primary importance was the publicity of the oath. Political documents that include oaths, such as inter-state treaties, were often put up in a polis sanctuary, but these decrees are unrelated to the particular workings of the sanctuary.56 Also, the oaths that are usually mentioned in Greek cult regulations are oaths of office of sacred personnel. Though these oaths may be sworn in sanctuaries, their causes are unrelated to judicial procedures or monetary values. A deme inscription from Cos of the first half of the third century BCE is a case in point. It states that the women who obtained the priesthood of Demeter by lot “should serve as priestesses after they swear an oath” (τὰς δὲ λαχούσας ὀμοσάσας ἱερῶσθαι).57 The sacred law from Andania dating to 91 BCE or 23 CE offers another good example for an oath of office of sacred personnel.58 The form and purpose of an oath mentioned in a fragmentary inscription from Erythrai (ca. 350–300 BCE) is unclear; the inscription states that the buyers of the priesthood of the Kyrbantes have the right to administer oaths to different people.59 The group includes priests and priestesses who have the right to act in place of the buyers as well as the people who have been initiated by them.60

• 61  For a cult regulation that stipulates that the priest and magistrates act together in response to (...)
• 62  I.Oropos 277.9–20 (= IG VII 235). For text and discussions, see also RO 27; CGRN 75.
• 63  For the two different kinds of cases, see Wilamowitz (1886), p. 94; I.Oropos 277, commentary p. 18 (...)
• 64  I.Oropos 277.16–17. The text also regulates how the priest should conduct the trials in the sanctu (...)
• 65  For the popularity of the sanctuary and its significance, see I.Oropos 277, commentary p. 182; Lup (...)

40Evidence that sanctuaries in ancient Greece exercised jurisdiction is almost nonexistent. An inscription from Oropos dating to 386–374 BCE offers unique insights into an otherwise unseen aspect of Greek sanctuaries.61 It describes a jurisdiction of the priest in the Amphiareion.62 The priest is given the right to issue fines up to the limit of five drachmae for offences by visitors against the sanctuary; he also acts as judge (δικάζει|ν, 13–14) in private disputes among the visitors up to the limit of 3 drachmae.63 Major offenses are referred to the respective common law courts.64 Just like the sanctuaries of Asclepius, the Amphiareion was a healing sanctuary visited by people in the hope that the god would appear to them while they slept in the temple. The Amphiareion was open for the major part of the year, and during this period it must have accommodated many visitors from different places. This particular situation, together with the fact that the sanctuary was located at a certain distance from the common courts of the polis, may explain why the local authorities granted the judicial authority to the priest in the first place.65 However, oaths are not mentioned in this document from Oropos, and it seems that the jurisdiction of the priest was a purely practical measure that only concerned minor offences happening in the sanctuary itself. It was a sort of extended policing.

• 66  For the corpus of manumission documents from Delphi, see SGDI II 1684–2342. In general, on the pra (...)
• 67  On this legal fiction (Scheinkauf), which is attested especially in Delphi, but also in a few othe (...)
• 68  SGDI II 2072 (198 BCE). Curiously, this inscription seems to be the only manumission document from (...)
• 69  SGDIII 2072.16 and 26, respectively. Though the ritual “at the altar” is not specified, it seems m (...)
• 70  SGDI II 2072.12–13.

41Inscriptions from Delphi dating to the Hellenistic period show that the sanctuary of Apollo played an important part in manumissions of slaves.66 The manumission took place in the sanctuary and involved the legal fiction that the owner sold his slave to the god.67 One of the documents concerns the manumission of two slaves.68 It specifies that the transaction took place “between the temple and the altar” (ταῦτα δὲ ἐγένετο ἀνάμεσον τοῦ ναοῦ καὶ τοῦ βωμοῦ) and that both the owner and the slaves sanctioned the conditions of the manumission by swearing the customary oath “at the altar” (ὤμοσαν ποτὶ τῶι βω{ι}μῶι).69 Among the witnesses are two priests and a temple warden.70 In this case, the sanctuary provides the legal service of a notary for a specific kind of commercial transaction.

• 71  For the corpus of the confession inscriptions, see Petzl (1994); Petzl (2019), p. 1–105. For the i (...)
• 72  See Varinlioglou (1989), p. 39, who highlights the fact that the “sacral procedure” (sakrales Verf (...)
• 73  For an overview of the thematic range of cases, see Petzl (1994), p. xii–xiv. For an analysis of t (...)
• 74  See Chaniotis (1997), p. 353–384, who points out that the use of judicial language in these texts (...)

42The so-called confession inscriptions (Beichtinschriften) also indicate a sort of temple jurisdiction.71 However, these much-discussed texts from a circumscribed rural area in Lydia dating to the second and third centuries CE are less concerned with actual trials, than with the avoidance of common justice and the notion of divine punishment.72 As personal dedications, the stelae tell the stories of people who turn to the temple, believing that a deity punished them with sickness for an offence committed in the past, and, after finding out the cause for the divine punishment, put up a stele in the sanctuary in order to demonstrate to others the power of the gods.73 While it is generally agreed that the sacred personnel of the local sanctuaries functioned as mediators between the individuals and the gods, it is a matter of debate to which extent the temples may have served as actual law courts exercising jurisdiction over the local population.74

• 75  In ancient Greece, this role of judicial oaths is usually associated with archaic and “primitive” (...)

43While these sources indicate that certain sanctuaries in Greece and Asia Minor could be involved in legal processes, whether for policing misconduct in the sanctuary (Oropos), or for certifying manumission contracts (Delphi), or for mediating divine justice (Lydia), ancient Near Eastern texts offer much more characteristic parallels for the oath regulations in the Marmarini inscription. In the cultural context of the ancient Near East, temples served as common courts of law, and evidentiary oaths played an eminent part as a means to decide cases.75 What is more, the Near Eastern cultural context offers evidence for the distinctive ritual oath gestures mentioned in the Marmarini inscription.

• 76  For the corpus of demotic temple oaths, see Kaplony-Heckel (1963); Massa (2018), who contextualize (...)
• 77  For overviews of the contents of the temple oaths, see Kaplony-Heckel (1963), p. 26–27; Massa (201 (...)
• 78  Kaplony-Heckel (1963), p. 258–259 (no. 154).

44The so-called Demotic temple oaths (Tempeleide) are a case in point. Hundreds of texts on ostraca and papyrus bear witness to the function of temples as common courts of law in Ptolemaic Egypt.76 These oath protocols show that the temple was a place to which the local population would travel when they wanted to decide personal disputes and quarrels over sales of goods, inheritances or other personal matters.77 A characteristic feature of the legal procedures is that all the cases are decided on the basis of evidentiary oaths: if a defendant swears the oath that is formulated for the particular case, he or she wins, but if the person decides not to swear, he or she loses. One of the temple oaths from Thebes, for instance, concerns the quarrel between two women over a loan.78 The plaintiff demands the return of a loan of 100 silver coins. The defendant claims that she did not receive the money. The ostracon gives the place and time of the oath; it mentions the wording of the oath; and it states the decision. This is the German translation by Ursula Kaplony-Heckel:

• 80  For the divine names and place names, see Kaplony-Heckel (1963), p. 25; Massa (2018), p. 118–119.

45According to this translation, the oath is taken “at (the gate) of the temple of Djeme, in the temple of Montu, lord of Medamud”.80 In the “decision”, the text states the principle that if the defendant swears that she did not receive the sum, the plaintiff should stop demanding the money; and if she does not take the oath, she should pay the money.

• 81  See Kaplony-Heckel (1963), p. 20–22; Massa (2018), p. 203–205. For a judicial oath in a temple, me (...)
• 82  Massa (2018), p. 205.
• 83  Massa (2018), p. 208.
• 84  See Massa (2018), p. 210–213, who notes that the symbolic gesture of grabbing earth mentioned in a (...)

46Several aspects of the Egyptian temple oaths are of particular interest for the present discussion: first, they are always taken in a temple, usually in the area defined as the gate or the dromos.81 As Viviana Massa has pointed out, these places are symbolically charged as “locations in hearing distance of the god”.82 Second, the judges are always priests or other officials associated with the temple.83 Third, the oaths are decisive for the outcome of the disputes of the local population; and fourth, the causes for the oath-trials are personal matters that mostly concern transactions of agricultural goods or money, that is claims that have a monetary value. It has been argued that the temple oaths probably involved oath gestures,84 but unfortunately the short formulaic texts do not provide information about this aspect.

• 85  In general, on ancient Near Eastern law, see Westbrook (2003); see also Walther (1917).
• 86  See e.g. Mercer (1912), p. 10–12; Dombradi (1996), p. 78–84, 368, 330–335 and 368; Faist (2015), p (...)
• 87  CT 45 37 (CDLI no. P285723); Westbrook (2009), p. 41–47 (for transliteration, translation and comm (...)
• 88  Westbrook (2009), p. 43.
• 89  See Westbrook (2009), p. 43–44. For the decisive role of an oath in a document dating to the reign (...)
• 90  See Westbrook (2009), p. 203; San Nicolò (1938), p. 309–310; Sandowicz (2012), p. 91–95, who point (...)

47Documents from Mesopotamia offer an even clearer link to the Marmarini inscription.85 Just as in Egypt, disputes between private individuals were usually decided in temples or at the gates of temples by the use of assertory oaths as case-deciding evidence.86 An Old Babylonian document, for instance, records a dispute over the validity of a sale of a female slave.87 It states that the judges passed judgement “in the temple of Shamash, in the Gate of the Shrine”.88 In this case, it was stipulated that the buyer declare under oath that she paid the full price. Though taking the oath would have decided the case in her favor, the woman chose not to swear. As a result, a compromise was reached and the plaintiff received one shekel.89 It has been pointed out that litigation in the temple was in use until the Neo-Babylonian and Persian periods and even gained in significance during these times.90

• 91  The use of περιλαμβάνω in the concrete sense “to grasp” is conspicuous. The verb is normally used (...)
• 92  See e.g. San Nicolò (1938), p. 311; Harris (1965), p. 217; Dombradi (1996), p. 78–86 and 332–334; (...)
• 93  See Walther (1917), p. 191–199, who cites numerous texts that mention either the “weapon” of diffe (...)
• 94  TS 58 (CDLI no. P498213); Charpin (1980), p. 242–243 (transliteration).

48Strikingly, the Mesopotamian documents show that an object defined as the “weapon” of a god played a crucial role in the trials. The Marmarini inscription mentions an oath gesture in which the oath-taker grasps “a weapon of the goddess” (ὅπλον τῆς θεοῦ).91 This gesture raises several questions: what exactly is the meaning of ὅπλον in this context? Where and how would the oath-taker grasp such an object? Why does the goddess have several weapons? And in what sense do these objects belong to her? While Greek sources offer no direct parallels that could help us to answer these questions, the Mesopotamian legal documents provide information that could almost be considered a template. It seems that it was common in Mesopotamian litigation that one of the litigants (usually the defendant) would swear an oath before a divine emblem.92 The person would take the oath while standing in front of the sacred object. Significantly, the divine emblem (šurinnu) is often called the “weapon” (giš-tukul/kakku) of the god.93 An Old Babylonian text from Tell Sifr (Kutalla), for example, concerns a dispute about the ownership of an inherited garden.94 It defines the place where the defendant swears the oath as a set-up of several divine emblems:

• 96  For the meanings of kakku, see CAD, above n. 93. The meanings of ὅπλον also include the special se (...)
• 97  See e.g. Eur. HF 161; 942; Xen. Hell. 5.1.12; 5.4.3; Plat. La. 183d; Aristot. De gen. an. 760b11; (...)
• 98  Eur. HF 942 (Heracles’ club); Call. Epigr. 1.7 Pfeiffer (staff of old man); Paus. Att. α 40 (aegis (...)
• 99  PBS VII, no. 85 (giš.tukul^dutu): Harris (1965), p. 219 (for translit., transl. and comment.).

49The range of meanings of the Akkadian word giš-tukul/kakku, which comprises “weapon” and “tool”, closely matches the meanings of ὅπλον.96 Just like the Akkadian word, the Greek word ὅπλον is not restricted to a specific kind of weapon; rather, it is a generic term denoting any implement or tool that has the potential of being used for fighting or is part of the armor.97 This includes typical military implements such as shields, swords and helmets, but also the club of Heracles, the staff of an old man, the aegis of Zeus or even an abstract idea like virtue.98 The “weapons” of the goddess in the Marmarini inscription may thus be either actual weapons or certain implements considered her divine weapons. Given the plurality of weapons of the anonymous goddess in the Marmarini inscription, it is also noteworthy that one Babylonian text refers to an oath sworn before two or more weapon-symbols of the god Shamash.99

• 100  For the so-called “journey of the divine weapon”, see Harris (1965), p. 217–224; Spaey (1993), p.  (...)
• 101  For this suggestion, see Pedersen (1914), p. 147, who refers to the general significance of touchi (...)
• 102  San Nicolò (1938), p. 308; Faist (2015), p. 66.
• 103  See Sandowicz (2011), p. 17–36, who shows that the fear of perjury was strong even in the neo-Baby (...)

50Further aspects appear relevant for our discussion. First, the Mesopotamian weapon-symbols of the gods are tangible, movable objects. In certain cases, these sacred objects would even be carried out of the temple to a field or another place of litigation and oath-taking.100 While the act of putting up a divine emblem in a sacred place is in itself a ritual action, there is also the possibility that the oath-taker would touch them.101 Second, the Mesopotamian texts do not mention standard oath formulas. It is generally understood that the wording of the oaths depended on the specifics of a case.102 This variability of the formulas, together with the emphasis both on the place of oath-taking (in the temple) and the ritually manipulated objects (the divine weapon-symbols), could explain why the oath regulations in the Marmarini inscription do not specify the wording of any of the oaths, either, but focus instead on ritual and place. Third, the Mesopotamian judicial oaths are, to a certain extent, voluntary oaths. The litigants are obliged to take them only if they want to win a case. Fearing the consequences of perjury—and it seems that the presence of the divine emblems served as a powerful reminder of the punitive power of the gods—, a person could choose not to take the oath and accept to lose the case.103 Such an element of choice is also implied in the oath section in the Marmarini inscription which opens with the words: “If someone takes an oath.” This formulation suggests a choice.

• 104  See CGRN 225, commentary, § 6 (lines A 48–54). For discussion, especially of the issue of the unsp (...)
• 105  See e.g. Berti (2006), p. 190 n. 6.
• 106  For discussion of the oath in the Critias, see Herter (1966), p. 236–259; Nesselrath (2006), p. 41 (...)
• 107  See above n. 41. For the touching of sacrificial pieces, see IG XII 4, 275.11–14 (Kos, mid-4th cen (...)
• 108  In the decree of the Demotionidai (IG II² 1237; RO 5; 396/5 BCE and after), the gesture of “holdin (...)

51The Marmarini inscription mentions another distinctive oath gesture, namely the action of drinking, either “from the altar” or “from the fish”. Both these actions have been described as “mysterious”.104 However, when they are seen as part of the variety of actions through which the idea of contact with the sphere of the gods could be expressed in ancient oaths, they start to make sense. The gesture in the Marmarini inscription has two significant ritual aspects. One aspect is the action of drinking (an unspecified liquid) itself, the other is the material object from which the oath-taker takes the ritual drink. It has been rightly pointed out that the act of drinking a liquid was not an established oath ritual in ancient Greece.105 The famous scene in Plato’s Critias (119c–120d) in which the mythical kings of Atlantis drink a mixture of wine and bull’s blood as part of a complex sacrificial oath-gesture that also involves a libation is the exception that proves the rule. Plato creates this oath-ritual for literary purposes, using both familiar and unfamiliar elements in order to highlight the importance of the meeting of the kings of Atlantis and to characterize them as divine descendants.106 As has already been noted, oath gestures in ancient Greece commonly revolved around the universal rituals of animal sacrifice and libation: the oath-taker would either pour a libation over a burnt sacrifice, or he would touch sacrificial pieces.107 In these cases, the ritual takes place at an altar or sacred stone. In very few cases, it seems, the altar is attested as an object the oath-taker would touch or from which he would take something, yet not a drink.108

• 109  On this much-discussed passage, see Levine (1993), p. 200–212; see also Fishbane (1974), p. 25–45; (...)
• 110  Levine (1993), p. 206–207, points out that in ancient Israel, the court of law would be located in (...)
• 111  Levine (1993), p. 210.

52Just as the role of a sanctuary as a place of jurisdiction and arbitration is a distinctive cultural feature of the ancient Near East, the use of the act of drinking as an oath ritual also points to this region. Comparative evidence suggests that it may have been common in the entire region. It is attested, for instance, in the Israelite legal practice. Num 5: 11–31, a famous passage known as “the ordeal of the errant wife”, concerns the situation that a man accuses his wife of adultery.109 The accusations are brought before the priest in the temple.110 The priest decides the case by administering an oath to the woman and making her undergo a ritual that has the following form: after mixing pure water with dust from the sanctuary and “inscribing” curses in it, the priest makes her drink “the bitter water causing the curse” (vehishka ‘et ha’isha ‘et-mey hamarim hame’ararim, 24). In his commentary, Baruch Levine compares the biblical oath ritual with an oath ritual in a letter from the royal palace of the Mesopotamian city-state Mari.111 In this letter, written in Old Babylonian, Queen Šiptu tells her husband King Zimri-Lim of a vision that had been reported to her and in which the gods swear an oath to protect the city:

• 113  VAT 9962 (CDLI no. P369111): Fishbane (1974), p. 39 (for transliteration and translation).

53The divine oath ritual described in this letter seems to have a basis in actual Mesopotamian oath practice. A fragmentary text dating to the middle Assyrian period mentions the act of drinking as an oath gesture in a litigation.113

• 114  KUB 13.4 (CTH 264, Middle Hittite period). For an edition, see Miller (2013), p. 244– 265, who hig (...)

54Hittite Anatolia provides perhaps the closest parallel for the ritual action mentioned in the Marmarini inscription. Among the Hittite documents pertaining to religion is an important text with cult regulations dating to the later Middle Hittite period that details the duties of temple personnel for the organization of the official cult in the temples and during festivals.114 The instructions concern, in particular, purity rules and provisions of vessels, food and animals. One passage addresses the situation that the cowherds and shepherds who are obliged to deliver animals to the temple may instead fancy other uses (§ 19):

• 117  Just like Miller (2013), Mouton (2017), p. 244, translates the phrase BI-IB-RU DINGIR-LÌ ZI-TI as (...)

56Though the Hittite document predates the Marmarini inscription by more than a thousand years, it offers a striking parallel for the ritual oath gesture in the Hellenistic inscription. First, it mentions the same action as in the inscription. In both cases, the oath ritual consists of an action defined as drinking from the altar. The Hittite text is, however, more specific; the two passages describe a scenario of how the ritual in the Marmarini inscription may have played out in practice. Second, the oath-taking in the Hittite text has a judicial context: by swearing it in the sanctuary, the shepherds and cowherds clear themselves from the suspicion that they may not have fulfilled their duties as required. Third, the Hittite text highlights the significance of the material aspect of the ritual by mentioning that the vessel that is used belongs to “the god himself/of life” (BI-IB-RU DINGIR-LÌ ZI-TI, § 18).117 In emphasizing this material aspect, rather than the liquid that was drunk, the text indicates that the association of the vessel with the deity is the main source for the religious power of the oath.

• 118  See CAD s.v. bibrû. For zoomorphic vessels in antiquity, including “rhyta” (defined in the strict (...)
• 119  See Koehl (2013), p. 238–247; see also Koehl (2018), p. 45–85, who points out (p. 51) that in the (...)
• 120  For animal-shaped drinking vessels or rhyta made for (often playful) use in the symposion in ancie (...)
• 121  For fish-shaped rhyta from Ugarit and Kamid el-Loz (both places in the Levant), see Koehl (2006), (...)

57There is another aspect that makes the Hittite text particularly interesting for our discussion. The word for the vessel mentioned in the text is the Akkadian term bibrû, which primarily means “bird”, but also refers to an animal-shaped vessel and is usually translated as “rhyton”.118 Scholars have shown that the bibrû is a zoomorphic drinking-vessel that is defined by its status as a cult object that can also represent a deity.119 Vessels in the shapes of various animals, such as bulls or rams, are known from various parts of the ancient Near East and Greece.120 Strikingly, however, fish-shaped vessels seem to have been a characteristic cultural feature of the ancient Near East, especially the Levant and Mesopotamia.121

• 122  Interestingly, the liquid is not specified. In view of the object (“the fish”), the oath context a (...)
• 123  For parallel passages and discussion, see Zerhoch (2025), p. 171 n. 53. For the meaning, see also (...)
• 124  For a reconstructed neo-Assyrian cultic stone water basin in Berlin (Vorderasiatisches Museum, Ide (...)
• 125  I.Smyrna 735 (LSAM 17; CGRN 245; the stele is now lost).

58Three scenarios emerge for the ritual oath gesture of drinking “from the fish” as mentioned in the Marmarini inscription; in all three, the relationship between the fish and the goddess and the idea of contact between the oath-taker and the goddess are crucial. The first scenario is that the fish is an animal-shaped vessel that represents the goddess. By drinking a liquid from such a sacred vessel, the oath-taker may feel to drink in a sense both from the same vessel as the goddess and “from the goddess”.122 This scenario would satisfy the meaning of the phrase “to drink from something” (πίνω ἀπό τινος), which usually denotes the place from which the liquid comes, for example a source of water, rather than the vessel with which one drinks the liquid.123 The second option would be that the phrase refers to an artistic/architectural feature connected to a water-source that is sacred to the goddess and from which the ritual drink would be taken.124 One may think, for example, of a fountain-head in the form of a fish. The third scenario would be that “the fish” is short for a fish-pond (ἰχθυοτρόφιον) in the sanctuary. A fish-pond is mentioned in a short sacred law from Smyrna dating to the first century BCE that corresponds closely with the inscription from Marmarini.125 The text may be cited in full:

• 127  For the anonymity of the main goddess in the Marmarini inscription, see Parker (2016), p. 58–59.
• 128  See I.Smyrna 735, commentary; see also CGRN 245, general commentary.
• 129  See Decourt – Tziaphalias (2015), p. 26–27; Parker (2016), p. 59; CGRN 225, general commentary.
• 130  For σκεῦος, see LSJ s.v.
• 131  For the meaning of ἰχθυόβρωτος, see I.Smyrna 735, commentary. For the religious prohibition to eat (...)

59As in the Marmarini inscription, the goddess remains unnamed in this document.127 Scholars proposed several identifications, including Atargatis, Cybele and Artemis (Ephesia or Anaitis).128 Two of these, Artemis and the Syrian goddess Atargatis, have been suggested as candidates for the unnamed goddess in the Marmarini inscription.129 Also, both texts highlight the significance of sacred implements belonging to the respective goddesses. The use of the word ὅπλον, “weapon, implement considered as a weapon”, in the Marmarini inscription (A 48, 49, 50), corresponds with the more general term σκεῦος, “implement, equipment of any kind”, in the inscription from Smyrna (2–3).130 In both texts, the unnamed goddess has several such weapons/implements. Though neither document specifies the implements, they are marked as possessions of a female deity and as objects used in cult. Moreover, both texts correspond with each other in that they refer to weapons/implements of a goddess and to fish in the same context. The Marmarini inscription speaks of “the fish” (τοῦ ἰχθύος, A 51) as a physical object representing the power of the goddess in the context of an oath gesture, thus implying a self-curse and divine punishment, while the inscription from Smyrna mentions fish as sacred animals (lines 1 and 8–9) and explicitly refers to fish as punitive instruments of the goddess in a conditional curse that is formulated as a warning for anybody who either harms the sacred fish or damages or steals the “equipment” belonging to the goddess: the text wishes for this person to become “fish food” (ἰχθυόβρωτος).131

• 132  For the lack of an archaeological context for the Marmarini inscription, see Decourt – Tziaphalias(...)

61First, the discussion has shown that the rubric on oath has a crucial impact on the understanding of the inscription as a whole. If the hypothesis is correct, the sanctuary functioned not only as a place of diverse worship of the anonymous goddess, but also exercised an important judicial function: it served as a small claims court. The oath section suggests that the visitors would bring their personal disputes to the unknown sanctuary132 so as to decide them there in the context of a legal practice based on evidentiary oaths. This practice suggests that the belief in the authority of the goddess as a source of truth was as strong as the belief in her punitive and protective powers. The emphasis on ritual oath gestures shows how these beliefs translated into the actual conflict management. The oaths received their particular power from ritual actions that signaled different levels of proximity to and contact with the goddess.

• 133  For a discussion on the question of which liquid may have been used, see also Zerhoch (2025), p. 1 (...)
• 134  As such, the act of drinking “from the fish” may be considered a rite de passage. For a fundamenta (...)

62Second, while the use of assertory judicial oaths is the main focus and organizing principle of the oath section—it concerns four of the five conditional clauses (A, B, C and E)—, there is also a rule for a promissory oath (D). As the proposed new reading in line A 51 suggests, this oath is linked to a specific task in the sanctuary, namely the office of the neokoros or temple warden. The common denominator of the five conditional clauses is the use of ritual oath-gestures implying the idea of proximity to the goddess. The ritual of the oath of the neokoros brings the person closest to the goddess. This ritual, which centers on the action of drinking “from the fish”, has two significant aspects. First, it makes use of fish as an object that represents the anonymous goddess herself; this use must have conveyed to the oath-taker a strong and tangible sense of the presence of the goddess. The Near Eastern texts in which the act of drinking appears as an oath gesture suggest that it may have been considered particularly powerful. Second, in view of the cultic importance of water and purification rituals in the inscription, and in view of the emphasis on ritual purity (hagneia) in this particular oath regulation (the person who wants to serve as neokoros is said to be/keep pure for his entire life), it seems most likely that the oath-taker would drink some sort of ‘sacred’ water “from the fish” and that this action signaled his obligation to strictly follow the purity rules.133 The gesture may also have indicated the commitment and new status of the person as sacred personnel.134

• 135  See Carbon (2016), p. 185–208 (festivals); Parker (2016), p. 58–59 (the anonymous goddess); Parker(...)

63Third, the oath section also offers important new indications for the larger question of the socio-cultural background of the inscription. Several studies have brought out the strong Near Eastern influence in this text.135 Significantly, the oath section points to a particular regional context within the Near Eastern sphere of influence. The legal practice in the sanctuary, the use of evidentiary oaths, the function of a small claims court and, above all, the distinct forms of the ritual gestures of grasping cult objects and drinking “from the altar” and “from the fish” all point to a Near Eastern cultural contact zone with a strong Mesopotamian and Anatolian influence, such as the Levant or Northern Syria. Given the fact that ritual oath gestures are culturally determined, the gestures mentioned in the Marmarini inscription offer more distinctive clues for its socio-cultural background than other rituals which are more universal. While “divine weapons” and the act of drinking are not part of ancient Greek legal practice or oath-taking in general, there is clear and distinct evidence that both were common in the ancient Near East.

• 136  See Decourt – Tziaphalias (2015), p. 45–46; Carbon (2016), p. 180–185; CGRN 225, general commentar (...)
• 137  The use of “oath fees” also supports the hypothesis of a private cult association.

64Though the Near Eastern sources are, of course, not direct sources of influence, they have a double value: they offer general indications for the possible socio-cultural background of the Marmarini inscription and provide possible scenarios for the meaning and purpose of the oath regulations. At the same time, the contextualization of the oath-regulations has made it clear that the religious life described in the Marmarini inscription is a unique mixture of influences. In addition to the Near Eastern sphere of influence for the legal-ritual practice, the oath regulations seem to make use of a formal legal principle that was common in ancient Greece. The use of the value of a claim as an organizational principle in legal practice is mentioned in several Greek sources. While in these sources different values mostly entail a distribution of the cases to different courts, that is to smaller or larger groups of judges, in the Marmarini inscription, they entail different forms of rituals and amounts of oath fees. In fact, the values of claims mentioned in the Marmarini inscription strongly suggest that the sanctuary served as a small claims court, deciding cases that may not be brought before the common courts. Furthermore, this judicial practice and oath-taking offers an indication that this sanctuary may have been located outside a polis and frequented by a community best described as local and, at the same time, multicultural, who strongly believed in the authority of oaths and the powers of the anonymous goddess. It has been suggested that a private association may have organized the life of the sanctuary rather than a polis.136 The oath regulations offer strong support for this assumption.137

• 138  A conclusive case for a specific identity of the anonymous goddess has not yet been made. For tent (...)
• 139  On Atargatis, see Hörig (1984), p. 1536–1581; Drijvers (1986), p. 355–358; Bilde (1990), p. 151–18 (...)
• 140  See Lightfoot (2003), p. 65–72 and 489–497. For fish as iconographic attributes of Atargatis, see (...)
• 141  Several scholars have pointed out the complexity of the situation; see e.g. Drijvers (1986), p. 35 (...)
• 142  Bilde (1990), p. 171.
• 143  On Allat, see Drijvers (1978), p. 331–358; Starcky (1981), p. 564–570; Linant de Bellefonds (2009) (...)
• 144  See Drijvers (1978), p. 331–358; Starcky (1981), p. 564–570.
• 145  See Kaizer (2013), p. 161–166, figs. 1–4 (for the association with fish). For two temple reliefs f (...)
• 146  For the so-called “Cerberus” relief from Hatra and the identification of the deities as Allat (usu (...)
• 147  Hdt. 1.131.3 and 3.8.3. For the etymology of the name “Allat”, see e.g. Starcky (1981), p. 564.

65The ritual of drinking “from the fish” also allows us to venture an outlook regarding the identity of the anonymous goddess in the inscription. Implying a physical object of some sort symbolizing the anonymous goddess and her power, be it a fish-shaped drinking vessel, an artistic/architectural feature or even a fish-basin, the rule that the oath-taker “should drink from the fish” provides a strong and distinctive argument that the anonymous goddess in the Marmarini inscription is indeed a Near Eastern goddess who may be seen more specifically as a universal goddess with origins in Syria and the Levant.138 The religious significance of fish in the region is well-known; it is usually associated with the name of Atargatis and the main cult center of this goddess at Hierapolis.139 Interestingly, the association of Atargatis with fish is not restricted to fish as sacred animals kept in a pond in the sanctuary, but extends to divine attributes or symbols, as ancient Near Eastern iconographic representations show.140 However, this does not mean that the anonymous goddess in the Marmarini inscription can be simply identified as Atargatis; the identification of female deities in the region is highly complex not only because the sources for most places are limited, especially in the Hellenistic period, but also because of cross-cultural influences and flexible divine interpretations.141 The cult of Atargatis spread widely in the entire region and beyond at a relatively early date, with the goddess developing into a universal divine figure “capable of several interpretations”.142 Atargatis shares this development with the Anatolian goddess Cybele, for instance, or the Arabian goddess Allat, who was also worshipped widely in the region, including the cities of Palmyra (Syria) and Hatra (Mesopotamia).143 In view of the oath regulations, the goddess Allat is of particular interest even though the iconographic and epigraphical evidence is younger than the inscription from Marmarini: Allat is often artistically represented with weapons and in armor as Athena, but also shown as a seated goddess who shares iconographic features with Cybele and Atargatis.144 Also, reliefs from Hatra not only depict her as a goddess of justice, but also associate her with fish.145 In one of the reliefs, Allat is connected with Nergal, a Mesopotamian god, who is mentioned in the Marmarini inscription as a Syrian interpretation of the Greek god Pan.146 What is more, her very name, which is first attested in Herodotus as Ἀλιλάτ, simply means “the goddess”.147